Designing Safer Nitriles - ACS Symposium Series (ACS Publications)

Chapter 10

Designing Safer Nitriles

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on August 27, 2015 | http://pubs.acs.org Publication Date: May 5, 1996 | doi: 10.1021/bk-1996-0640.ch010

Stephen C. DeVito Office of Pollution Prevention and Toxics, U.S. Environmental Protection Agency, Mail Code 7406, 401 M Street, SW, Washington, DC 20460

Nitriles represent an important class of chemical substances that have broad commercial utility. A comprehensive search and review of the literature to identify studies pertaining to toxic effects caused by nitriles revealed that certain nitriles are acutely toxic (lethal) or may produce osteolathyrism. A retrospective analysis of these studies, particularly those that describe toxic mechanisms and provide structure-activity relationship data, enabled an understanding of why certain nitriles are highly toxic while others are not. From this understanding, structural modifications that reduce toxicity became apparent. This paper describes how safer (less toxic) nitriles can be designed. Chemists who design nitriles will find this paper useful for designing safer, commercially efficacious nitriles. The general approach and strategy described herein for the design of safer nitriles can be used for the design of safer substances belonging to other chemical classes as well.

The design of a safer substance begins with the concomitant consideration of structural features that are needed for commercial usefulness along with any toxicological effects that are caused by such features. In addition, the toxic effects that can be caused by other structural features intended to be present in a planned substance needs to be considered as well. Chemists developing a particular type or class of substance for a specific purpose (e.g., dye, surfactant) are aware of the structural features that are necessary for use function, and will design substances such that they contain these structural features. In many instances, however, chemists are probably not aware of the toxic effects associated with the structural

0097-6156/96/0640-0194$17.50/0 © 1996 American Chemical Society

In Designing Safer Chemicals; DeVito, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.


10.

DEVITO

Designing Safer Nitrites

195

features or the chemical class to which a planned new chemical belongs. Consequently, the toxic properties are likely to be retained in the new substance. If a structural feature that is necessary for commercial use is likely to bestow toxicity, the chemist should consider making molecular modifications that w i l l reduce the toxicity but will not affect the substance's commercial usefulness. But how does one know when a planned substance is likely to be toxic? The potential for a planned substance to be toxic may be inferred from what is known about other chemicals that contain similar structural features. Whether one is striving to design a new, less toxic analog of an existing substance, or develop a series of substances for a particular commercial purpose, it is imperative that the chemist be aware of the toxicity-related information available for the existing substance, or the class to which the series of substances belongs. To design a safer chemical, the need to identify and familiarize oneself with any toxic properties and toxicity data related to the chemical class is just as important as a familiarity with the chemistry of use. How does one identify such information? Moreover, how can one use such information to design safer chemicals? Chapter 2 provides a list of reference sources that are extremely useful for identifying toxicity-related information. Chapter 2 also describes several ways in which safer chemicals can be designed and, using many examples, demonstrates and provides insight into how one can infer, from available information, structural modifications that can reduce toxicity without affecting commercial efficacy. Data that pertain to mechanism of toxicity, metabolism, and structure-activity relationships for a series of analogous compounds are particularly useful to assess the toxicity of a new or planned substance, and for inferring structural modifications that will reduce toxicity. Using nitriles as an example of an important commercial class of substances, the present chapter answers the above questions more fully, and demonstrates how safer chemicals can be designed from an analysis of the toxicity-related information available in the literature. More specifically, this paper will demonstrate the importance of the following principles of designing safer chemicals: • the need to first familiarize oneself with any toxicological or health related data available for existing analogous substances (i.e. substances belonging to the same class as the intended safer chemical); • when toxicological effects are known for the chemical class to which a planned new substance belongs, how these toxic effects can be minimized by ultilizing available toxicological data (e.g., toxic mechanisms, structure-activity data, metabolism data, etc.) to infer structural modifications that will reduce or prevent the same toxicity in the new substance.


196

DESIGNING SAFER CHEMICALS

The approach described herein for the design of safer nitriles can be applied to the design of safer chemicals of any chemical class.


Nitriles Nitriles are organic substances that contain the cyano ( C = N ) group. Nitriles have wide commercial application, that includes their use as solvents, synthetic intermediates, pharmaceuticals, and monomers, to name a few. To establish if any toxic effects are associated with nitriles, a literature search was conducted to identify information pertaining to toxic properties of nitriles. This began with a search of the Chemical Abstracts Database, starting with the first Decennial Subject Index (1907), and searching each subsequent decennial subject index. Using the subject heading of "nitriles", the decennial indices were searched for literature citations pertaining to toxicity or health-related studies of nitriles. In addition, toxicological textbooks (1,2) were searched similarly. These reference sources were also searched for toxicity-related data pertaining to specific, commercially important nitriles (e.g., acetonitrile, propionitrile) because it was felt that studies on these nitriles may also contain data on other nitriles as well. The literature search revealed that much has been published on the toxicity of nitriles. Literature references containing toxicity information on nitriles were obtained and analyzed. It was found that, as a class, there are two toxic properties associated with nitriles. These are acute lethality and osteolathyrism. Some nitriles possess both of these toxic properties. Each of these properties are discussed separately below. Acute Lethality of Nitriles Certain nitriles are quite potent in causing acute lethality in humans and animals (3,4). Consequently, a considerable number of studies have been conducted to determine the acute lethality of a variety of nitriles used commercially (3-28). Table I summarizes some of the available acute lethality data measured during these studies, and pertains specifically to acute lethality measurements made i n mice following oral administration of each nitrile listed. Acute lethality is expressed as the acute median lethal dose (LD50). The lower the L D 5 0 value, the more toxic the substance. It can be seen from Table I that as a class nitriles vary broadly in their ability to cause acute lethality, and that subtle differences in structure can dramatically affect toxic potency. For example acetonitrile (1) is not very acutely lethal (LD50 = 6.55 mmol/kg) whereas its homolog propionitrile (2) is ten-fold greater in toxicity (LD50 = 0.65 mmol/kg). Increasing the carbon chain length by one more methylene ( - C H 2 - ) group (i.e., butyronitrile, 3) however, only slightly increases toxicity. In fact, the toxicity decreases sharply as the number of - C H 2 - groups increases above two, as can be seen in comparing the L D 5 0 values of 3-7. It is also


In Designing Safer Chemicals; DeVito, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996. 2

788.9 515.2 300.0 12703.4 34808.9 703.9 70.2

2

2

5

3

4

5

2

2

2

2

2

2

2

2

3

2 2

2

2

2

7

6

4

3

2

2

2

2

2

3

2

2

2

kaporr*

500.2 0.005 0.041 12703.4 34808.9 3.3 0.82

4.0 251.0 58.4 24.7 16.9 8.1 5.5 8288.5 0.98 6.6 4288.2 5.4 28.9 41.7 113.5 202.1 107.0 4.9 16.2

(x 10-24)

d

15.30h (-1.185) 9.19 (-0.963) 19.40* (-1.288) 13.37 (-1.126) 27.50i (-1.439) 17.99 (-1.255) LOOT ( 0.000) 0.34 ( 0.474) 0.39f ( 0.409) 0.25 ( 0.598) 0.89f ( 0.051) 3.03 (-0.481) 48.72f (-1.688) 10.62 (-1.026)

3.42 (-0.534) 0.89 (0.051) 1.15 (-0.059) 1.54 (-0.188) 2.25 (-0.352) 8.71 (-0.940) 23.28 (-1.367) 0.34 (0.470) 3.31 (-0.520) 2.61 (-0.416) 0.39 ( 0.398) 7.59 (-0.880) 3.13 (-0.495) 3.34 (-0.524) 1.53 (-0.185) 0.95 ( 0.023) 1.07 (-0.029) 2.32 (-0.366) 1.64 (-0.213)

calcd e

(log(I/LD50))

6.55f (-0.816) 0.65*" (0.187) 0.57f (0.244) 2.30'(-0.362) 4.77f (-0.679) 14.09f (-1.149) 14.79*" (-1.170) 0.37f (0.432) 2.80f (-0.447) 5.02f (-0.701) 0.29f (0.538) 1.80e (-0.255) 1.62u (-0.210) 2.83e (-0.452) 1.59B (-0.201) 1.03s (-0.013) 1.84f (-0.265) 0.57f (0.244) 0.52f (0.284)

exp

LD50

° Octanol/water partition coefficient. * Theoretical reaction rate constants for cytochrome P450-mediated oc-hydrogen atom abstraction (ref. 22). c Theoretical reaction rate constants statistically corrected for metabolism at other positions (ref. 22). d Acute oral median lethal dose in mice (mmol/kg). * Calculated by eq. 1. f Obtained from ref. 11. g Obtainedfromref. 13. h Obtainedfromref. 3. *" Obtainedfromref. 29.

5

6

2

2

3

3

2

2

3

6

2

3

2

2

3

3

2

2

3

3

2

2

3

3

CH (CH ) CN CH (CH ) CN CH (CH ) CN CH (CH ) CN CH (CH ) CN (CH ) CHCN (CHsbCHC^CN (CH ) CH(CH ) CN CH CH CH(CH )CN NCCH CN NC(CH ) CN NC(CH ) CN NC(CH ) CN NC(CH ) CN C1CH CN C1(CH ) CN C1(CH ) CN

(CH ) N(CH ) CN -0.24 (CH ) CHNH(CH ) CN 0.12 -0.92 HN(CH CH CN) 0.12 CH =CHCH CN C H CH CN 1.56 C H CH CH CN 1.55 -1.10 HO(CH )2CN

2

a

k *>

(x 10-24) 4.0 287.9 310.0 324.9 3249 348.8 348.8 8422.2 97.4 349.9 581L7 5.4 28.9 70.6 226.9 495.0 107.0 45.8 13L1

logP« -0.39 0.14 0.66 1.19 1.72 2.78 3.31 0.44 1.06 1.59 0.97 -1.20 -0.82 -0.96 -0.43 0.11 0.22 0.20 0.38

CH3CN CH3CH2CN

compound no. structure

acetonitrile 1 propionitrile 2 butyronitrile 3 valeronitrile 4 capronitrile 5 caprylonitrile 6 pelargonitrile 7 isobutyronitrile 8 3-methylbutyronitrile 9 4-methylvaleronitrile 10 2-methylbutyronitrile 11 malononitrile 12 succinonitrile 13 glutaronitrile 14 adiponitrile 15 pimelonitrile 16 chloroacetonitrile 17 3-chloropropionitrile 18 4-chlorobutyronitrile 19 propionitrile, 3-(dimethylamino)- 20 3-(isopropylamino)- 21 3,3'-iminodi22 3-butenenitrile 23 21 phenylacetonitrile 3-phenylpropionitrile 25 3-hydroxypropionitrile 26

name

Table L Mouse Acute Toxicity Data, Log P, and Theoretical Metabolic Rate Constants of Some Nitriles


198



interesting to note the influence of methyl substitution on toxicity. Methyl substitution of butyronitrile (3) in the 3-position greatly reduces toxicity, as can be seen when comparing the toxicity of 3 to that 9. However, methyl substitution of butyronitrile in the 2-position nearly doubles toxicity, as evident in comparing the toxicity of 3 to that of 11. Mechanism of Acute Lethality. It has been observed that exposure of humans and experimental animals to the more acutely toxic nitriles results in signs and symptoms similar to that of cyanide poisoning, implicating free cyanide as the cause of lethality (3,4)- Lang observed in 1894 that cyanide is in fact released from nitriles following their administration to mammals: large amounts of thiocyanate ( a metabolite of cyanide) appeared in the urine of dogs and rabbits following oral administration of several nitriles (5). Subsequent studies by other investigators have shown that cyanide is released from a wide variety of nitriles in other species and is associated with the acute lethality of these substances (6-23). Ohkawa and co-workers proposed that cyanide release from nitriles results from the metabolism of these substances in the liver. They proposed that cyanide liberation from nitriles results from cytochrome P450-mediated hydroxylation of the carbon atom alpha (a) to the cyano ( C = N ) group to form a cyanohydrin intermediate, which rapidly decomposes to liberate cyanide and the corresponding carbonyl compound (7). Subsequent studies have confirmed the role of hepatic cytochrome P450 enzymes in the release of cyanide from various nitriles following the administration of these substances to laboratory animals (7-27). Generally speaking, nitriles that are metabolized most readily or most quickly at the carbon atom alpha to the cyano group (i.e. the a carbon) are more toxic than nitriles that are metabolized more slowly at this position (22, 22). Thus, the mechanism for the acute lethality of nitriles is directly related to their propensity to release cyanide as result of metabolic bioactivation by cytochrome P450 enzymes. As discussed elsewhere in this book (see chapter by Jones), radical formation is a requisite step in cytochrome P450 hydroxylations, and it is now known that cytochrome P450-mediated hydroxylation proceeds by a stepwise process involving radical substrate intermediates (28-20). With respect to the toxicity of nitriles, this would involve radical formation on the acarbon, followed by subsequent hydroxylation to the cyanohydrin intermediate. Cytochrome P450-mediated release of cyanide from nitriles is depicted in Figure 1. Not surprisingly, nitriles which liberate cyanide more quickly are more toxic. The less toxic nitriles (the ones which have mouse oral L D 5 0 values larger than about 2 mmol/kg) are less toxic because they tend to release cyanide more slowly and to a much lesser extent than nitriles that have L D 5 0 values below 2 mmol/kg (11,12). It is important to emphasize



10.

DEVTTO


199

that cytochrome P450-mediated hydroxylation can occur at other carbon positions (i.e. postions that are not alpha to the cyano group) as well. Hydroxylation at positions other than the a-carbon does not result in cyanide release and represents a detoxication pathway. In fact, it is likely that most nitriles, even the toxic ones, are hydroxylated at multiple carbon positions. The toxic nitriles, however, are those in which metabolism at the a-carbon predominates, and the less toxic nitriles are those in which metabolism at non-a-carbon positions predominate (21,22). This point is discussed in greater detail in the next section and elsewhere in this chapter. Structure-Activity Relationships of Acute Lethality. The structureactivity relationships of the acute toxicity of nitriles have been reviewed(22). Because the acute toxicity of nitriles is a function of their ability to undergo cytochrome P450-mediated hydroxylation on the carbon atom alpha to the cyano group, and that the hydroxylation is a radicalbased reaction (Figure 1), the acute toxicity of nitriles is expected to relate to structural features that influence a-carbon radical stability. Logically, structural features which are expected to increase a-carbon radical stability are likely to favor a-hydrogen atom abstraction. The more quickly a - h y d r o g e n atom abstraction occurs, the more quickly cyanohydrin formation occurs and the more quickly cyanide is released. The more quickly cyanide is released, the more toxic the nitrile is expected to be. In examining the toxicity data in Table I one can observe a direct relationship between acute toxicity and the relative ease of radical formation (hydrogen atom abstraction) on the a-carbon. It appears that the differences in the toxicities between the nitriles in Table I can be ascribed to the type of radical that can be formed at the a-carbon. For example, the a-carbons of isobutyronitrile (8) and 2-methylbutyronitrile (11) are tertiary (3°) and, of course, will form 3° radicals at the a-carbon. These nitriles are considerably more toxic than propionitrile (2) and butyronitrile (3), which form less stable secondary (2°) radicals on their a carbon atoms. Propionitrile, on the other hand, is considerably more toxic than acetonitrile (1) because the latter substance forms a less stable primary (1°) radical at the a-carbon. The large difference in toxicity between chloroacetonitrile (17) and 1 is likely to be due to the fact that the chloro substituent of 17 is expected to promote cyanohydrin formation because it will stabilize the radical formed at the a-carbon (23), and thereby favor its formation, when compared to 1. Phenylacetonitrile (24) contains benzylic hydrogens at the a-carbon. These hydrogens are particularly easy for cytochrome P450 enzymes to remove because the resultant carbon radical is benzylic, and well stabilized through resonance of the phenyl ring. In fact, this substance is one of the most toxic of all the nitriles listed in Table I. In Designing Safer Chemicals; DeVito, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.


200


Thus, the differences in the acute toxicities of the nitriles in Table I are attributable to the most likely position of cytochrome P450-mediated metabolism. The carbon atoms most likely to be hydroxylated by cytochrome P450 are those which will form the more stable carbon radical. While structural features that favor radical formation at the occarbon tend to increase toxicity, it also appears that structural features that favor radical formation at other carbon atoms decrease acute toxicity. For example, the a-carbons of nitriles 3, 4, 9 and 10 are secondary, but 9 and 10 are considerably less toxic than 3 and 4, respectively. This is because nitriles 9 and 10 contain tertiary carbon atoms elsewhere in their structure, and cytochrome P450-mediated hydroxylation is expected to occur preferentially at these tertiary carbon atoms rather than the secondary carbons. Hydroxylation at the tertiary carbon of 9 or 10 will not result in cyanide release. For nitriles with multiple C H 2 (i.e. secondary) carbons, the large decrease in toxicity observed as the number of C H 2 groups increases (e.g., nitriles 3-7) is likely to be due to the fact that these extra methylene units offer additional sites of metabolism that compete for metabolism at the a carbon. It has been shown that release of cyanide from nitriles 4-7, and 910 is very slow (11,12) and is likely to be the reason why these nitriles are considerably less acutely toxic than nitriles such as 2-3, 8, 11, and 24, which release cyanide much more quickly (i.e. are metabolized predominately at the cc-carbon) (11,12). A similar point can be made in comparing the extreme differences in toxicities of propionitrile (2) to 3-hydroxypropionitrile (26). Placement of the hydroxy group onto the terminal carbon of propionitrile (to give 26) results in a very large decrease in toxicity (Table I). Although the products of metabolism of 26 have not been fully studied, one investigation did not detect any cyanide in the blood of rats administered 26 at a dose of 3 g/kg (route not specified) (24). It is known that alcoholic (C-OH) carbons are easily metabolized (25) and it seems likely, therefore, that metabolism of 26 occurs predominately (if not entirely) at the C-OH carbon. Thus, the 3-hydroxy group of 26 directs cytochrome P450 mediated metabolism away from the a-carbon, thereby reducing the acute toxicity of this substance relative to 2. It is noteworthy that there is a very large difference in toxicity between 26 and its isomer 2-hydroxypropionitrile (27), shown on the next page. The rat oral L D 5 0 of 27 is 1.23 mmol/kg (28). (Mouse oral L D 5 0 data for this substance are not available). On the other hand, 26 has very low toxicity: the rat oral LD50 of 26 is 45 mmol/kg (2). The reason for the high toxicity of 27 is because this substance is a cyanohydrin, the bioactivated form (Figure 1), and as such decomposes readily to release cyanide. Similar to hydroxy-substituted nitriles, the toxicity of aminosubstituted nitriles can vary widely, depending upon the location of the amino group. Acute toxicity data for some a-nitriles and |}-amino


10.

DEVTTO

201

Designing Safer Nitriles CH —CH —C^J 2

2

CH3—CH—ON OH

OH

26


rat oral L D

5 0

27 = 45 mmol/kg

rat oral L D

5 0

= 1.23 mmol/kg

nitriles are presented in Table I and Table II. The p-amino nitriles shown in Tables I and n are derived from propionitrile. The toxicity data in Table I were measured in mice, whereas the data in Table II were measured in rats. It is readily apparent that the p-aminopropionitriles (20-22, 32, 33) are considerably less acutely toxic than propionitrile (2). The low toxicity of 20-22,32 and 33 compared to 2 may be explained from the influence of the amino substituents directing metabolism away from the carbon atom alpha to the cyano group. Mumtaz et al. reported that in rats approximately 44% of orally administered 3-(dimethylamino)propionitrile (20) is excreted unchanged, and 3-aminopropionitrile and cyanoacetic acid are the only urinary metabolites (26). Keiser et al. reported that 3-aminopropionitrile is almost completely excreted in the urine as cyanoacetic acid when administered to mice or rats (27). These studies clearly indicate that beta (P)-amino nitriles are metabolized at the carbon atom adjacent to the amino group, and not the carbon alpha to the cyano group. Thus, it appears that cyanide is not released to any large extent from P-amino nitriles. However, nitriles substituted with an amino group at the carbon alpha to the cyano group are quite toxic (Table II). The acute toxicity of both aminoacetonitrile (28) and 2-(dimethylamino)acetonitrile (29), for example, are greater than the acute toxicity of acetonitrile (1) by approximately 100-fold (Table II). Similarly, the acute toxicities of the butyronitrile-derived a-amino nitriles 30 and 31 are substantially greater than that of butyronitrile (3). Thus, the presence of an a-amino group in a nitrile greatly increases acute toxicity. Unlike the p-amino nitriles described above, metabolism data for a - a m i n o nitriles 28-31 are unavailable and consequently the reason for their extreme toxicity is not as readily apparent. What is particularly interesting, however, is that 30 and 31 contain a-methyl groups as well: no a-hydrogen atoms are present. In addition, the 3-position of 31 contains a tertiary hydrogen atom. These structural characteristics of 30 and 31 are expected to decrease acute toxicity if cytochrome P450-mediated a-hydroxylation were the basis of toxicity. The reason for the high toxicity of a-amino nitriles may be due to their chemical similarity to cyanohydrins: the amino group of an a amino nitrile may be behaving similarly to the a-hydroxy group of a cyanohydrin, and as such is expected to release cyanide rapidly without


202


H R l

_ I_ f R

cyto. P-450 _H»

N

• R

l

R

2

nitrile Ri=R , H, alkyl or aryl

cyto. P-450 hydroxylation

f

R

2

2

cyanohydrin intermediate

alpha carbon radical


2

O

acute lethality

II

HCN

C

/ \ cyanide carbonyl metabolite

Figure 1. Mechanism of Acute Toxicity of Nitriles: Cytochrome P450Mediated Cyanide Release. Table II. Acute Toxicity (LD ) Data of Some Alpha (a)-, and Beta (P)-Amino Nitriles 50

Oral LDso (mmol/kg)

compound no.

acetonitrile* propionitrile* butyronitrile«

structure

mice

rats

CH3CN CH3CH2CN CH3CH2CH2CN

6.55* 0.65* 0.56&

78c 3.26c 3.18c

a-Aminonitriles aminoacetonitrile 2-(dimethylamino)acetonitrile 2-amino-2-methylbutyronitrile 2- amino-2 3-dimethylbutyronitrile

28 29 30 31

NH CH CN (CH3)2NCH CN CH3CH C(CH3)(NH2)CN CH3CH(CH3)C(CH )(NH2)CN

p-Aminonitriles 3- aminopropionitrile 3-methylaminopropionitrile 3-(dimethylamino)propionitrile S^'-iminodipropionitrile

32 33 20 22

NH CH CH CN

4.3*

CH3NHCH2CH2CN

41.6c

/

2

2

2

2

3

2

2

2

(CH3)2NCH CH CN HN(CH CH CN) 2

2

a

2

2

2

15.3c 27.5*

c

0.47 1°. The presence of a hydroxy or a substituted or unsubstituted amino group on the a-carbon atom greatly increases toxicity, whereas the presence of these substituents at other carbon positions greatly reduces acute toxicity. Quantification of Structure-Acute Lethality Relationships. As discussed in Chapter 2, quantification of structure-activity relationships enables precise prediction of biological activity (toxicological or pharmacological) directly from structure. Quantitative structure-activity relationships are particularly useful for the design of safer chemicals because one can quantitatively predict the toxicity of untested or planned substances, and observe the influence of structural differences between analogous substances on toxic potency. Lipophilicity is an important factor of biological activity of a substance because it is a descriptor of bioavailability (this is discussed in Chapter 2). For this reason log P (octanol/water partition coefficient; a physicochemical expression of lipophilicity) is widely used, either alone or in combination with other physicochemical properties, as a descriptor in the quantification of structure-activity relationships. Several investigators have attempted to quantify the structure-acute lethality relationships of nitriles by correlating into a regression equation L D 5 0 data with certain physicochemical descriptors (11-13, 21, 22). Tanii and Hashimoto (11-13) had limited success in correlating the L D 5 0 data of many of the nitriles in Table I with log P. It was found that log P was not a good overall descriptor of acute lethality of nitriles: they observed that the acute lethality of certain nitriles correlate well with log P, whereas others do not. Those nitriles in which acute lethality correlates with log P only correlate when divided into subgroups; each subgroup having its own regression equation. These regression equations are very limited for predicting the acute toxicity of untested nitriles because how is one to know if the nitrile whose toxicity is to be predicted is one in which acute



204


lethality correlates with lipophilicity. Even if this where known, it is not clear which of the regression equations is the most appropriate to use. What is more useful is a single regression equation that accurately quantifies the structure-acute lethality relationships of a large group of structurally diverse nitriles. A possible reason why the acute lethality data of nitriles does not correlate well in a single regression equation in which log P is the only term may be because the log P term is not a descriptor of the mechanism of acute lethality of nitriles. From Table I one can readily see that there are nitriles that have similar lipophilicity but differ greatly in acute lethality. Phenylacetonitrile (24), for example, has essentially the same lipophilicity as 10 but is over 10-fold more acutely toxic. The same is true in comparing 12 with 26,11 with 9, or 11 with 4. The differences in the toxicities of these substances is related to their differences in the type of carbon radical formed at the carbon atom alpha to the cyano group following cytochrome P450-mediated hydrogen atom abstraction (metabolism). From the L D 5 0 data of the nitriles in Table I, Grogan and co-workers (22) developed a single quantitative structure-activity relationship equation (equation 1) that uses log P and a descriptor (kanr) of cytochrome P450-mediated hydrogen atom abstraction at the a-carbon, relative to hydrogen atom abstraction at other carbons. l o g ( l / L D ) = -0.16(logP)2 + 0.22(logP) + O.ll(lnfc^orr) + 5.67 50

n = 26

(1)

r = 0.85

In equation 1 n is the number of nitriles used and r is the correlation coefficient. The k^ort term is derived from k , which is a descriptor of cytochrome P450-mediated hydrogen atom abstraction at the a-carbon only and does not consider cytochrome P450-mediated hydrogen atom abstraction at other carbon atoms. The k term is calculated from ionization potential, ground-state and radical heats of formation (22). The k corr term takes into account the likelihood of cytochrome P450mediated hydrogen atom abstraction occurring at all non a-carbon atoms relative to the likelihood of hydrogen atom abstraction at the a-carbon, and is a much better descriptor of overall cytochrome P450 metabolism than is k . The acute toxicity data of the nitriles in Table I correlated well with equation 1 (22). a

a

a

a

Table I contains the k and k rr values for each nitrile used in the derivation of equation 1. The meaningfulness of these descriptors with respect to acute toxicity of nitriles becomes apparent upon examination of the data provided in Table I. Nitriles whose k values are considerably higher than their corresponding k ^ values (e.g. nitriles 4-7,10,20-22) are those that are metabolized predominately at non-alpha carbon positions and, consequently, are less toxic. Nitriles with high k r values (e.g. a

aCO

a

aCor


10.

DEVTTO


205


nitriles 8,11, 23, 24) are metabolized predominately at the a-carbon, and are quite toxic. Thus, these descriptors give an indication of the extent of cytochrome P450 hydrogen atom abstraction that is expected to occur at the a-carbon relative to other carbons in the molecule. For an untested or planned nitrile, these descriptors (as well as log P) can be derived directly from structure (22) and, using equation 1, one can estimate the L D 5 0 value (acute toxicity) of the nitrile. The Design of Less Acutely Toxic Nitriles. How can the information discussed in the preceding paragraphs be used to design nitriles that are less acutely toxic? The above analysis of the literature sources pertaining to the acute toxicity of nitriles provides insight into how less acutely toxic nitriles can be designed. More specifically, the studies that have explored and elucidated the mechanism of acute lethality enable a clearer understanding of the relationship between structure and acute lethality. Guidelines for the design of less acutely toxic nitriles are provided below. When designing a nitrile, add structural features that will prevent or minimize cytochrome P450-mediated hydroxylation at the a-carbon. Remember that the acute toxicity of a nitrile is directly related to structural factors that influence the ease in which cytochrome P450 enzymes catalyze radical formation at the a-carbon atom relative to the other carbon atoms within the substance. Acute toxicity is enhanced when the a-carbon can form a stable radical, such as when the a-carbon is tertiary or benzylic. Therefore, one should avoid having tertiary or benzylic a-carbons in a nitrile. If the a-carbon must be tertiary or benzylic for purposes of intended use, add other substituents that compete with, or direct cytochrome P450 metabolism away from the a-carbon. This could be accomplished by making another (i.e. non-a) carbon in the nitrile benzylic, tertiary, or alcoholic (i.e., add an O H group). Or, if allowable from a use standpoint, design a nitrile that has a quarternary a carbon (i.e., an a-carbon that is bonded to three other carbon atoms, in addition to the cyano group). Such a nitrile should have low acute toxicity because it does not contain an a-hydrogen and thus cannot form a cyanohydrin intermediate upon cytochrome P450 metabolism. Unless absolutely necessary for intended use, do not add hydroxy or amino groups (substituted or unsubstituted) to the a-carbon, because these nitriles will be highly acutely toxic. As discussed earlier, nitriles that are cyanohydrins or contain an amino group (substituted or unsubstituted) on the a-carbon are highly acutely toxic because they release cyanide readily, without the need for bioactivation from cytochrome P450. Because such nitriles do not require cytochrome P450bioactivation to release cyanide, substituents added to redirect metabolism


206



away from the a-carbon will have little, if any, effect in reducing acute toxicity. There is very little that can be done to reduce the toxicity of a cyanohydrin or a-amino nitrile except, perhaps, to incorporate structural modifications intended to minimize absorption (see Chapter 2). Calculate the log P, ka and kapm values for a planned nitrile and , using equation 1, estimate its LD50 value. Details for calculating log P, ka and kaoon and using equation 1 are provided in reference 22. Equation 1 will be very helpful in assessing the acute toxicity of a planned nitrile or a series of nitriles, and for assessing the effects of structural modifications on acute toxicity. When considering structural modifications that are intended to reduce acute toxicity (lethality), be careful not to choose and incorporate structural modifications that will reduce acute toxicity, but will bestow other toxicity, such as osteolathyrism or neurotoxicity. In addition to acute lethality, osteolathyrism is another other health risk posed by certain nitriles. At least two nitriles are known to be neurotoxic. One needs to be careful not to make molecular modifications that will reduce acute lethality, but will make a nitrile an osteolathyrogen or a neurotoxicant. For example, addition of a dimethylamino group to the 3position of propionitrile (2), to give 20, results in a substantial decrease in acute toxicity; 20 is substantially less acute toxic than 2 (Tables I and II). However, 20 is highly neurotoxic whereas 2 is not. Thus, although addition of the dimethylamino group to the terminal carbon of 2 greatly reduces the acute lethality of 2, this structural modification bestows neurotoxicity. A discussion of the nitriles that cause osteolathyrism and neurotoxicity is provided below.

Nitrile-Induced Osteolathyrism Osteolathyrism is a medical condition characterized by hernias, dissecting aortic aneurysms, lameness, skeletal deformities such as exostoses and kyphoscoliosis, and slowing or cessation of body growth. These medical complications are, of course, quite serious and often lead to premature death. Osteolathyrism occurs when there is interference with the ability of collagen and elastin to undergo the crosslinking that normally occurs with aging. Thus, the condition is more likely to occur (and is more serious) during the early stages of growth and development (e.g., childhood), where collagen and elastin have not yet fully crosslinked, than it is during later stages of life, when the majority of crosslinking of collagen and elastin has already occurred. Striking characteristics of osteolathyrism in laboratory animals include: weakening of the arteries, tendinous and ligamentous attachments, epiphyseal plates, skin, cartilage; delayed wound healing; and marked deformations of the skeleton (35-37). Osteolathyrism is



10.

DEVITO


207

caused by exposure to certain chemical substances during early stages of life, particularly during gestation, infancy and early childhood. Certain hydrazines, hydrazides, semicarbazides and nitriles, for example, are known to cause osteolathyrism. Several comprehensive reviews on osteolathyrism and the chemical substances that cause it are available (3840). The first indication that nitriles can induce osteolathyrism came when the condition was observed in animals that ingested seeds of the leguminous plants of the genus Lathyrus , which includes many kinds of peas (38). A crystalline substance producing the signs and symptoms of osteolathyrism was isolated from Lathyrus Odoratus almost simutaneously by Schilling and Strong (41) and Dasler (42). Using degradation and unambiguous synthesis, Schilling and Strong (43) later identified this substance to be b-(N-g-L-glutamyl)-aminopropionitrile, of which the osteolathyritic principle is 3-aminopropionitrile (32). Diets containing the sweet pea Lathyrus Odoratus or 32 produce the signs and symptoms of osteolathyrism in several animal species (35, 44-52). Although 3-aminopropionitrile-induced osteolathyrism has not, to date, been reported in humans, the ability of this substance to induce this condition in many species strongly suggests that humans exposed to this substance may be at risk of developing osteolathyrism. The literature reveals that many other nitriles have been tested for inducing osteolathyrism in a variety of mammalian and nonmammalian species (35,38,47,53-58). These nitriles are listed in Table i n . Also shown in Table III is a relative ranking of the potency of these nitriles to induce osteolathyrism. This relative ranking was based on a retrospective analysis of the experimental results of the above studies. As can be seen from Table HI, of the many nitriles tested, only a few are known to produce osteolathyrism. Of these, 32 is by far the most studied with respect to causing osteolathyrism (35,38,47-73). Mechanism of Nitrile-Induced Osteolathyrism. The majority of studies conducted to elucidate the mechanism of nitrile-induced osteolathyrism have utilized 32 and, to a lesser extent, aminoacetonitrile (28) (54, 74-77). The early studies that utilized 32 found that the osteolathyrism induced by this substance resulted from an interference with intermolecular crosslinking within collagen and elastin fibrils (59-63). To enable a better understanding of the mechanism of nitrile-induced osteolathyrism, a brief discussion of collagen and elastin physiology is provided below. Collagen is the principal protein of connective tissue. Its major function is to provide strength and maintain the structural integrity of tissues and organs. Collagen fibrils are made up of tropocollagen subunits that are composed of three polypeptide chains. The tropocollagen subunits are connected to each other by covalently crosslinked lysine or 5hydroxylysine residues resulting from the action of lysine oxidase, a monoamine oxidase requiring copper and pyridoxal phosphate as cofactors. The crosslinking increases with age. Elastin, like collagen, is


208



Table III. Summary of the Results of Studies that have Tested Nitriles for Osteolathyrism «

relative potency in causing osteolathyrism*

compound name

acetonitrile propionitrile butyronitrile malononitrile succinonitrile phenylacetonitrile 3-hydroxypropionitrile aminoacetonitrile 2-aminopropionitrile 3-aminopropionitrile 3-(methylamino)propionitrile 3-(dimethylamino)propionitrile 3-amino-2-methylpropionitrile 3,3-iminodipropionitrile 3 3 3"-iminotripropionitrile 4-aminobutyronitrile ethylaminodipropionitrile benzylaminopropionitrile phenylethylaminopropionitrile 3,3'-oxydipropionitrile S^'-thiodipropionitrile I

/

/

no.

structure

1 2 3 12 13 24 26 28 34 32 33 20 35 22 36 37 38 39 40 41 42

CH3CN CH3CH2CN

CH CH(NH )CN NH CH CH CN

0 0 0 0 0 0 0 +++ 0 +++

CH3NHCH2CH2CN (CH )2NCH CH2CN

0/+c 0

NH CH CH(CH3)CN HN(CH CH CN)2 N(CH CH CN)

0/+d 0

CH CH CH CN NCCH CN NCCH CH CN C6H CH CN HOCH CH CN 3

2

2

2

2

2

5

2

2

2

NH2CH2CN 3

2

2

2

2

3

2

2

+++

2

2

2

2

2

3

NH2CH2CH2CH2CN

CH3CH N(CH2CH CN)2 2

2

C6H5CH2NHCH2CH2CN

C H5(CH2)2NHCH2CH CN 0(CH CH CN)2 S(CH CH CN) 6

2

2

2

2

2

2

0 0 0 0 0 0

* Results were obtained from refs. 35,38,47,53-58, and reflect effects observed in at least one species. & Relative potency is based on an analysis of the data presented in these studies, and are strictly qualitative. The scale used here with respect to causing osteolathyrism is as follows: 0 = not found to cause osteolathyrism; + = weak; ++ = moderately potent; +++ = highly potent, c Experimental results are ambiguous. Most studies show that this substance is non-osteolathyrogenic. Experimental results are ambigous. Some studies show this substance to be negative for osteolathyrism, while other studies show that it is positive. d



10.

DEVITO


209

rich in glycine and alanine but, unlike collagen, contains few prolines and consequently has elastic properties. It is present in ligaments and the walls of major arteries. The basic subunit of elastin fibrils is tropoelastin. Tropoelastin subunits are connected by the crosslinking of lysine residues, resulting from the action of lysine oxidase. Covalent crosslinking i n collagen and elastin takes place almost entirely through amino and/or aldehyde groups. The aldehyde groups are formed by lysine oxidase-catalyzed oxidative deamination of the epsilon (e) amino group of lysine or 5-hydroxylysine. The crosslinks occur by either imine formation between the aldehyde and amino group or by aldol condensation. The metabolic turnover of collagen and elastin is greatest during the developmental and growth periods of mammals, and gradually becomes very low i n adult life. As an animal matures the imine moieties are modified, presumably by reduction or condensation with other tropocollagen or tropoelastin subunits. The crossslinks become chemically more stable and the collagen or elastin fibers become increasingly rigid with age (78-92). A decrease i n crosslinks, as i n osteolathyrism, leads to an increase i n collagen and elastin fragility and, eventually, the signs and symptoms of osteolathyrism (discussed earlier). In mice and rats, the osteolathyrogenic effects of 32 occur after chronic administration rather than acute administration, and predominantly i n the weanling rather than the adult mouse or rat (27, 50, 64). Several studies found that 28 produced a very pronounced osteolathyrogenic effect i n the fetuses of pregnant rats (76, 35, 47). It had the strongest effect on the fetal skeleton and cardiovascular system. Collectively, these studies strongly indicate that the osteolathyrogenic actions of 28 and 32 result primarily from the inhibition of new covalent crosslink formation rather than by rupturing previously formed crosslinks. Inhibition of Lysine Oxidase. Page and Benditt concluded that 3aminopropionitrile-induced osteolathyrism results from the inhibition of oxidative deamination of lysyl residues, and not from direct binding of 32 to collagen (68, 70). Other investigators subsequently reported that 32 irreversibly inhibits lysine oxidase (69,71,76,77). It is now well established that the osteolathyritic properties of 28 and 32 are due to the ability of these substances to irreversibly inhibit lysine oxidase (56, 58, 77, 93). It is believed that the osteolathrytic properties of 22, 33, and 35 are also due to lysine oxidase inhibition (40, 58). As discussed above, lysine oxidase is a highly important enzyme because it functions during the initial phase of collagen and elastin formation. Its inhibition effectively results i n the complete blockade of collagen and elastin biosynthesis, and accounts for the extreme toxicity of substances that cause osteolathyrism from its inhibition. Several studies have been undertaken to elucidate the specific mechanism by which 28 and 32 irreversibly inhibit lysine oxidase (73, 76, 93). Maycock and co-workers (76) reported that the mechanism of action of enzymes that oxidize amines to aldehydes (e.g., lysine oxidase)


210


involves enzyme-mediated hydrogen atom abstraction from the carbon atom that is oxidized (i.e. attached to the amine). Thus, the presence of at least one hydrogen atom on the carbon that is attached to the amine is required for the oxidation. They found that aminoacetonitrile inactivates rabbit plasma amine oxidase in vivo. The compound showed no substrate activity but was a potent irreversible inhibitor (76. Experiments with [l- C]aminoacetonitrile showed that the enzyme became covalently labeled (76). The irreversible inhibition is thought to result from proton abstraction of the a-carbon which leads to the formation of a reactive ketenimine (Figure 2). The ketenimine intermediate, which is highly electrophilic, most likely undergoes attack from a nucleophile located within the enzyme active site resulting i n covalent bond formation and inactivation of the enzyme. Rando (73) suggested a similar mechanism for the inhibition of lysine oxidase by 32. Rando proposed that the amino group of 32 reacts with the aldehyde moiety of pyridoxal phosphate (a cofactor for lysine oxidase) to form an imine. Enzyme-mediated hydrogen abstraction from the - C H 2 - group adjacent to the imino group liberates a cyanide ion and yields a reactive imino-enamine product. It was postulated that the imino-enamine product undergoes a 1,4-addition reaction with some nucleophile located within the enzyme active site. In a later study, Tang and co-workers (93) found that cyanide is not liberated during the inhibition of lysine oxidase by 32, and proposed a slightly different mechanism of inhibition that is shown i n Figure 3. They proposed that following imine formation of 32 with pyridoxal phosphate a hydrogen atom from the carbon atom attached to the cyano group is removed by the enzyme, yielding a ketenimine intermediate (Figure 3). This ketenimine intermediate reacts as described i n the preceding paragraph, resulting i n irreversible inhibition of lysine oxidase. It seems plausible that irreversible inhibition of lysine oxidase by 28 may also involve initial imine formation with pyridoxal phosphate. Although it is generally accepted that imine formation between the amino group of 32 and pyridoxal phosphate occurs during the lysine oxidase inhibition cascade, it is not clear as to whether it is a requisite step for ketenimine formation (56, 58,93) . The amino groups of nitriles such as 3-(methylamino)propionitrile (33) and 3,3'iminodipropionitrile (22) are substituted and cannot form imine adducts with pyridoxal phosphate, yet these substances have been reported to cause osteolathyrism (Table III). The osteolathyritic effects of 22 and 33, however, are somewhat ambiguous. Some studies have found these substances to be inactive as osteolathyrogens. In addition, of these two substances, only 22 inhibits lysine oxidase in vivo (i.e., i n the body) and the inhibition is largely reversible (58). Clearly, the more potent amino nitriles are those which contain an unsubstituted amino group and, hence, can form imines. It is also noteworthy that both 22 and 33 are known to be metabolized to 3-aminopropionitrile (32) (58), and this may explain or at least contribute to their osteolathyrogenic properties observed i n some studies. Thus, although it is not entirely clear as to whether initial imine formation is essential for lysine


14


10.


DEVTTO

/ Nu

LysO \ Nu

LysO / \ + Nu Nu

c

H -

g-proton abstraction

H N — C — CEN 2

211

I

H?N— C = C==

H N — C—- C=N 2

N

H

aminoacetonitrile

ketenimine


28 Nu'

.LysO Nu H

I

I

H N— C — C = 2

I

NH

irreversibly inhibited lysine oxidase

Figure 2. Mechanism of Irreversible Inhibition of Lysine Oxidase (LysO) by Aminoacetonitrile, 28 (adapted from Maycock, et. al, ref. 76). / Nu

O II

CH

^CsN CH

2

H O

\ /

H +

C H z —

—

°

f

-

H

\

I

11

J

J

H N - C — C — CeN

0

II

HQ

C H

NH

1 -Y^Y

LysO \ Nu

I

I

H

H

H

2

pyridoxal-5-phosphate

2

32

imine adduct

a-hydrogen atom abstraction

LysO / \ + Nu Nu

Nu—LysO H

+ I

H

I

Nu

II

I

H

LysO \

+ Nu

I

•C=

NH

H N - C — C HQ

/

I

HN-C—C=C=

II

HC

I

H

I

>

Nu

N H N - C — C — C^N

H

irreversibly inhibited lysine oxidase

H

I

I

HC

H

H

o

ketenimine

Figure 3. Mechanism of Irreversible Inhibition of Lysine Oxidase (LysO) by 3-Aminopropionitrile, 32 (adapted from refs. 73 and 93).


212



oxidase inhibition, it at least appears to have considerable significance with regard to osteolathyrogenic potency. Structure-Activity (Osteolathyricity) Relationships of Nitriles. As can be seen from Table HI, only a few nitriles are known to cause osteolathyrism. Unlike the relatively large amount of single specie acute toxicity data available for nitriles (Table I), the osteolathyritic data compiled in Table III is comparatively limited for discerning structure-osteolathyricity relationships. This is because only a few nitriles are known to cause osteolathyrism and because the data used to derive the relative potencies in Table III are largely qualitiative and were collected in a variety of species. Nonetheless, a retrospective analysis of the osteolathyricity data available for nitriles and other substances, in conjunction with the mechanism of nitrile-induced osteolathyrism, provides some insight into the relationships between nitrile structure and osteolathyricity. This insight can be used for assessing the likelihood of a planned or untested nitrile having osteolathyritic properties, and for the design of nitriles that will not cause osteolathyrism. The general requirements for a nitrile to cause osteolathyrism are a cyano group and an amino group. Nitriles that do not contain an amino group are not osteolathyritic. Acetonitrile (1) and propionitrile (2) for example do not produce osteolathyrism, whereas aminoacetonitrile (28) and 3-aminopropionitrile (32) are potent osteolathyrogens (Table III). Amines that do not contain a cyano group, such as methylamine (CH3NH2), also do not cause osteolathyrism (47,54). Thus, both an amino group and a cyano group are essential requirements for a nitrile to cause osteolathyrism, and are consistent with the proposed mechanisms for lysine oxidase inhibition. It is also interesting to point out that 3hydroxypropionitrile (26), the hydroxy isostere of 32, is not osteolathyritic. This also is consistent with the mechanism proposed for irreversible inhibition of lysine oxidase by 32, in that 26 cannot form an imine. Another important general requirement for a nitrile to cause osteolathyrism is the presence of at least one hydrogen atom on the carbon atom adjacent to the cyano group. This requirement is inferred from the mechanisms of nitrile-induced inhibition of lysine oxidase (Figures 2 and 3). Nitriles that do not contain hydrogen atoms at the carbon atom adjacent to the cyano group cannot form lysine oxidasecatalyzed ketenimine intermediates and, logically, should not cause osteolathyrism. The distance between the amino and cyano groups appears to be important for osteolathyritic activity. While 28 and 32 are potent osteolathyrogens, 4-aminobutyronitrile (37) does not produce osteolathyrism. Thus, osteolathyritic activity is lost when the number of carbon atoms separating the cyano group from the amino group is three. Presumably, amino nitriles in which the number of carbon atoms separating the cyano group from the amino group is greater than three



10.

DEVTTO


213

will also not cause osteolathyrism. Experiments need to be conducted to confirm this presumption. The position of the amino group also appears to be important with respect to osteolathyritic activity. With the exception of 37, nitriles that have an amino group on their terminal carbon (e.g., 28 or 32) are potent osetolathyrogens, whereas osteolathyrogenic activity is lost when the amino group is located on a non-terminal carbon. This can be seen, for example, in comparing the osteolathyritic potency of 32 to that of 34; 32 is highly potent whereas 34 is inactive. The reason for these differences this is unclear. It is also not clear from the available data whether this is a function of the amino group being non-terminal or its distance from the cyano group. The amino groups of 28,32 and 34 are primary and, thus, capable of forming imine intermediates with pyridoxal phosphate. In addition, each of these nitriles contain at least one hydrogen atom of the carbon atom adjacent to the cyano group, and are expected to be removable by lysine oxidase to form a ketenimine intermediate. One would expect, therefore, that 34 and 37 would be capable of causing osteolathyrism. The apparent inability of 34 and 37 to cause osteolathyrism is not explainable from present knowledge, but could be related to steric factors. Substitution of the amino group greatly influences osteolathyritic potency. The most potent osteolathyritic nitriles are those that contain a primary amino (i.e., N H 2 ) group on their terminal carbon. Nitriles that have either a tertiary amino group (e.g., 20, 36, and 38), or a secondary amino group that contains a large subtituent (e.g., 39, 40) do not cause osteolathyrism. Nitriles that have a secondary amino group that contains a relatively small substituent, such as seen in 33 or 22, may still be osteolathyritic, but are less potent. These observations are generally consistent with the mechanisms proposed for nitrile-induced inhibition of osteolathyrism shown in Figures 2 and 3. The nitriles that contain primary amino group can undergo imine formation with pyridoxal phosphate, whereas tertiary and and secondary nitriles cannot. As discussed earlier, it is not entirely clear if initial imine formation is required for lysine oxidase inhibition but it certainly appears that nitriles that can form imine intermeditates are potent osteolathyrogens (with the exception of nitriles 34 and 37, which are not osteolathyrogens for other reasons) compared to nitriles such as 22 and 33, which cannot form imine intermediates. As stated previously, in laboratory animals nitriles 22 and 33 and known to be metabolized to the potent osteolathyrogen 32, and this may explain or at least contribute to their observed osteolathyritic properties. It is interesting to note that nitriles 41 and 42, the oxygen and thio isosteres (respectively) of 22 cannot be metabolized to 32 and are not osteolathyrogenic (Table HI). The Design of Non-Osteolathyritic (Safer) Nitriles. Just as an understanding of the structure-activity and mechanistic data pertaining to the acute toxicity of nitriles enables one to infer structural modifications



214


that can be used for the design of less acutely toxic nitriles, so can an understanding of the structure-activity and mechanistic data pertaining to the osteolathyricity of nitriles enable one to infer structural modifications for the design of nitriles that w i l l not cause osteolathyrism. When designing a nitrile, the only time one needs to be concerned about the possibility of osteolathyrism is when the nitrile will contain an amino group. Nitriles that do not contain an amino group are not osteolathyritic. Therefore, unless absolutely necessary, it is best not to incorporate an amino group onto a nitrile. In cases where a nitrile must contain an amino group, particularly a primary amino group, it is advisable that the amino group not be on a terminal carbon atom because at this position the amino group appears to bestow substantial osteolathyritic potency. If an amino group must be on a terminal carbon, there should be at least three carbon atoms that separate the amino group from the cyano group. Alternatively, one may also want to consider making the amino group a secondary amine that contains a large substituent (e.g., phenyl, benzyl), or a tertiary amine. Any of these structural characteristics are expected to minimize the potential for osteolathyrism. Placement of an amino group on a non-terminal carbon is not expected to bestow osteolathyritic properties. However, an amino group should never be placed on the carbon atom that is adjacent (alpha) to the cyano group because, as discussed earlier, such amino nitriles are highly acutely toxic. Another structural characteristic that should mitigate the risk of osteolathyrism of an amino nitrile is the lack of hydrogen atoms on the carbon atom alpha to the cyano group. Because the mechanism by which amino nitriles induce osteolathyrism is believed to involve hydrogen atom abstraction at the carbon atom alpha to the cyano group (Figures 2 and 3), it is logical to assume that nitriles that do not contain hydrogen atoms at this position will not cause osteolathyrism. Replacement of the alpha hydrogens of osteolathyritic amino nitriles with, for example, methyl or ethyl groups should eliminate osteolathyritic potency. Neurotoxic Nitriles The literature search for identifying toxic properties of nitriles revealed that two nitriles, 3-(dimethylamino)propionitrile (20) and 3,3iminodipropionitrile (22), are neurotoxic. The neurotoxic signs and symptoms exhibited by 20 and 22 are distinct and different from one another, suggesting that their neurotoxic mechanisms are different. There are some nitriles that exhibit neurotoxicity in laboratory animals, but only when administered in high doses or under conditions that could only exist in a laboratory (94,95). Except under conditions of extremely high exposure, it would seem unlikely that such nitriles would produce neurotoxicity in humans. Because only two nitriles exhibit neurotoxicity following exposure concentrations that are known or likely to be encountered in such places as the workplace, neurotoxicity is not a health


10.

DEVITO


215


concern for nitriles in general. This section summarizes the neurotoxic properties of 20 and 22, and suggests structural modifications that may attenuate their neurotoxicity. Neurotoxicity of 3-(Dimethylamino)propionitrile (20). The neurotoxic effects of 3-dimethylaminopropionitrile were first realized when an epidemic of urinary retention among workers exposed to the substance occurred in the spring of 1978 (96). The outbreak occurred in a plant that manufactured automobile seat cushions from polyurethane foam that contained a catalyst which consisted of 20 in 95% concentration. Workers experienced difficulty in urinating and lost any sensation of bladder distention. Exposure to the catalyst was through direct skin contact with foam cushions and by inhalation of catalyst vapor released from the foam following curing. Initially, worker complaints were limited to urinary retention. However, the affected workers later complained of loss of sensory acuity in the lower half of the body, parasthesia, muscle weakness, and sexual dysfunction (97). The urinary retention in these patients was attributed to effects on the peripheral nervous system. The inner lining (mucosa) of their bladders were normal (97). Of eight patients who underwent neurological testing during recovery, seven had a subclinical sensory abnormality and lacked either detrusor reflex or normal sensation of bladder filling; three had prolonged sacral-evoked responses; and two of these three had limb motor neuropathies (98). With cessation of use of the catalyst improvement occurred rapidly in nearly all cases, but ceased after one year. Repeat examinations two years later showed that those people with the most severe symptomatology had persistant urological, sexual and neurological abnormalities (99). Based on these observations and a subsequent animal study, it was concluded that 20 was responsible for the neurotoxic effects in these workers (100). Jaeger and Plugge evaluated the acute urinary bladder toxicity of 20 in male rats (101). It was observed that large doses rapidly produced central nervous system excitation that was followed by depression and death. Single oral doses of 0.31 or 0.62 mL/kg induced acute urinary bladder lesions in three days. Bladder changes consisted of massive transmural edema, acute ulcers and inflammation in essentially all animals, and hemorrhagic necrosis of the bladder wall in some of the animals tested. Jaeger and Plugge suggested that the bladder lesions may have been caused by the metabolite cyanoacetic acid. The ability of 20 to induce bladder damage in rats is probably not related to the urinary retention observed during occupational exposure, since workers exhibiting this effect had normal bladder mucosa as previously stated. The neurotoxic mechanism of 20 has not yet been elucidated. Although 20 has been found to inhibit acetylcholinesterase (102) and the glycolytic enzymes glyceraldehyde-3-phosphate dehydrogenase and phosphofructokinase (103), inhibition of these enzymes would not explain the urinary retention and the other neurotoxic effects caused by 20. This substance does not inhibit monoamine oxidase (104). Several authors have postulated that the neurotoxic effects of 20 are related to its



216


metabolism. Ahmed and Farooqui have reported the presence of cyanoacetaldehyde and cyanoacetic acid in the urine of animals administered 20, and suggested these metabolites are responsible for the neurotoxicity and bladder irritation, respectively (105, 106). It seems unlikely, however, that these metabolites play a role in the neurotoxicity or bladder irritation of 20. Both 3-aminopropionitrile (32) and 3-(methylamino)propionitrile (33) are metabolized to cyanoacetaldehyde and cyanoacetic acid, yet neither 32 or 33 are neurotoxic or produce bladder irritation (104). Thus, the neurotoxic mechanism of 20 remains an interesting enigma. Because the mechanism of neurotoxicity of 20 is unknown, and structure-activity (neurotoxicity) data are limited, it is difficult to infer a safer analog of 20. A possibly safer analog of 20 is its homolog 4(dimethylamino)butyronitrile. This substance does not produce urinary retention in rats (102). Further studies need to be undertaken to assess more fully any neurotoxic properties of this substance. No neurotoxic effects have been reported for 2-(dimethylamino)acetonitrile (29), the other homolog of 20. However this substance is not a safer substitute for 20 because it is considerably more acutely toxic (Table II). Neurotoxicity of 3,3'-Iminodipropionitrile (22). The neurotoxic properties of 3,3'-iminodipropionitrile (22) were originally reported by Bachhuber and co-workers (47). These investigators observed that rats fed diets containing 0.3% 22 developed muscle spasticity, weakness of the extremities, unsteady gait, and turned in circles. Similar neurotoxic effects have been reported in mice and birds (107). Over the years, other neurotoxic that include hearing loss, and memory and learning deficits have been reported in laboratory animals (108-115). No similarities to 3-(dimethylamino)propionitrile (20)-induced neurotoxicity have been reported, and the neurotoxic properties of 20 and 22 appear to be unrelated. The mechanism of neurotoxicity of 22 has been explored by Sayre and co-workers (116, 117). The neurotoxicity of 22 is believed to involve covalent bond formation of a metabolite of 22 with neurofilament fibers, resulting in the rapid dissociation of these fibers from their neuromicrotubules and subsequent neurotoxicity. These investigators postulate that 22 undergoes N-hydroxylation via flavin monooxygenase to N-hydroxy-3,3'-iminodipropionitrile (43) which, through a series of steps, is converted to the imine 44 (Figure 4). This latter substance undergoes tautomerization to dehydroiminodipropionitrile (45). It is believed that 45 undergoes direct attack by nucleophiles (e.g., an N H 2 group) located within the neurofilament fibers (Figure 4a). Alternatively, 45 may react with water to yield cyanoacetaldehyde (46) and 32 (Figure 4b). The former substance can react covalently with neurofilament fibers. The end result, in either case, are covalently modified neurofilament fibers and neurotoxicity. Of the two possible means of covalent modification neurofilament fibers (Figure 4a or 4b), the reaction of 45 with neurofilaments seems more plausible because substances such as 32 and 33 are also metabolized to cyanoacetaldehyde but these substances are not neurotoxic. If metabolism to



10.


DEVITO

FMO

H

217

OH | — N s C C H C H N C H2C HgCsN 2

2

—-

N=CCH CH N=ChCH C=N 2

2

2

N H C C H C H N CHgCHgCsN 2

2

43

44

22

covalently bound neurofiliment fiber

H -NCH=CK>=N

neuro- s filiment d—NH fiber >

2

N s C C H C H N C H = CHC=N 2

2

45 b

neurotoxicity

0=CH3H C=N 2

/ HzNCHzCHzCsN

46

32

Figure 4. Proposed Mechanism of Neurotoxicity of S^'-Iminodipropionitrile, 22 (adapted from refs. 116 and 117). FMO is flavin monooxygenase enzymes.



218


cyanoacetaldehyde were the basis of neurotoxicity caused by 22, then 32 and 33 would be expected to be neurotoxic as well. Assuming that the neurotoxic mechanism of 22 proposed by Sayre, et al. is correct, an approach that could be taken to design of a safer analog of 22 could be to incorporate structural modifications that prevent formation of 45. Since formation of 45 involves loss of a hydrogen atom at either the 3- or 3'-carbon atom, replacement of the hydrogen atoms at these positions with, for example, methyl groups will prevent formation of 45 and, hence, the resultant neurotoxicity. An analog such as 3,3'-tetramethyl-3-3'-iminodipropionitrile (47), for example, is not expected to be neurotoxic. CH

CH

3

3

N ^ C — C H — C — N — C — C H — ChN 2

2

I

I

CH3

CH3

47 This substance is also not expected to cause osteolathyrism for reasons discussed previously. It is possible however, that 47 may be significantly acutely lethal because in this substance the carbon atoms attached to the cyano groups are secondary and more likely to undergo cytochrome P450-mediated hydroxylation (which would result in cyanohydrin formation and subsequent release of cyanide) than are the CH3 carbon atoms. Equation 1 could be used to estimate the acute lethality 47. It is noteworthy to mention that the oxygen and sulfur isosteres of 22 (41 and 42, respectively) are not neurotoxic (38). The inability of 41 or 42 to cause neurotoxicity may be ascribed to their inability to undergo flavin monooxygenase-mediated hydroxylation to products analogous to 43, and supports the mechanism of neurotoxicity of 22 proposed by Sayre and co-workers (Figure 4). Neither 41 or 42 cause osteolathyrism (Table III), and are probably not acutely toxic (their oral L D 5 0 values are likely to be similar to that of 22). Thus both 41 and 42 are safer analogs of 22. The use of isosteric substitution in the design of safer chemicals is described in Chapter 2. H N^CCHCH2NCH2CH2C^N

22

2

NsCCHCHOCHCHC=N

N=CCHCHS CHCHC=N

41

42

2

2

2

2

2

2

2

2


10.

DEVTTO


219


Conclusion The literature search for identifying toxic effects caused by nitriles revealed that there are two general health concerns for this class of substances: acute toxicity (lethality) and osteolathyrism. Neurotoxicity is known to result from two nitriles. Chemists who design nitriles need to be aware of the possibility that these toxic effects may be caused by new nitriles, unless structural modifications that are intended to reduce or prevent toxicity are included during the design of new nitriles. A retrospective analysis of the literature studies has enabled the deduction of general structural modifications that prevent or reduce toxicity. Chemists who design nitriles can use these general structural modifications for the design new nitriles that are commercially efficacious and of low toxicity. One important caveat is that great care must be taken not to incorporate structural modifications that reduce or prevent one particular toxic effect but bestow another. This is why it is important to identify all toxic endpoints of concern, and understand, to the extent possible, any mechanistic and structure-activity data. A n understanding of such information will greatly minimize the likelihood of incorporating structural modifications that reduce one particular type of toxic effect but cause another. The general strategy described herein for determining structural modifications for the design of safer nitriles can be applied to other chemical classes as well. The important lesson to be learned from this chapter is that before designing chemical substances, chemists need to first familiarize themselves with all toxic properties that are associated with the chemical class to which the substance(s) under development belongs. Studies which describe or propose a mechanistic basis of toxicity are particularly useful for clarifying structure-activity relationships, and inferring structural modifications that mitigate toxicity. Once armed with such knowledge, the environmentally conscious chemist is in a much better position to design less toxic substances. Disclaimer This chapter was prepared by Dr. Stephen C. DeVito in his private capacity. The contents of this chapter do not necessarily reflect the views, rules or policies of the U.S. Environmental Protection Agency, nor does mention of any chemical substance necessarily constitute endorsement or recommendation for use. Literature Cited 1. 2.

Cassarett and Doull's Toxicology, the Basic Science of Poisons, Fifth Edition. Klaassen, C.D., Ed; McGraw-Hill: New York, 1996. Patty's Industrial Hygiene and Toxicology, 3rd Revised Edition, Wiley-Interscience, New York, 1982.


220



3.

Zeller, Von H . ; Hofmann, H. Th.; Theiss, A . M . ; Hey, W. Zentralbl. Arbeitsschutz 1969, 19, pp 225-238. 4. Hartung, R. Cyanide and Nitriles. In Patty's Industrial Hygiene and Toxicology, Clayton, G.D.; Clayton, F.E., Eds.; 3rd Revised Ed., WileyInterscience, New York, 1982; Vol. 2C (Toxicology), pp 4845-4900. 5. Lang, S. Arch. Exp. Path. Pharmak 1894, 34, pp 247-258. 6. Stern, J.; Weil-Malherbe, H.; Green, R.H. Biochem. J. 1952, 52, pp 114-125. 7. Ohkawa, H . ; Ohkawa, R.; Yamamoto, I.; Casida, J.E. Pest. Biochem. Physiol 1972, 2, pp 95-112. 8. Willhite, C.C.; Smith, R.P. Toxicol Appl. Pharmacol. 1981, 59, pp 589-602. 9. Silver, E.H.; Kuttab, S.H.; Hasan, T.; Hassan, M . Drug Metab. Disp. 1982,10, pp 495-498. 10. Ahmed, A.A.; Farooqui, M.Y.H. Toxicol. Lett. 1982, 12, pp 157-163. 11. Tanii, H.; Hashimoto, K. Arch. Toxicol. 1984, 55, pp 47-54. 12. Tanii, H.; Hashimoto, K. Toxicol. Lett. 1984, 22, pp 267-272. 13. Tanii, H.; Hashimoto, K. Arch. Toxicol. 1985, 57, pp 88-93. 14. Johannsen, F.R.; Levinskas, G.J. Fundam. Appl. Toxicol. 1986, 7, pp 690-697. 15. Yoshikawa, H . Med. Biol. 1968, 77, pp 1-4. 16. Davis, R.H. Cyanide detoxification in the domestic fowl. In Cyanide in Biology ; Vennesland, B.; Conn, E.E.; Knowles, C.J.; Westley, J.; Wissing, F. Eds.; Academic Press, London, 1981, pp 51-60. 17. Froines, J.R.; Postlethwait, E.M.; LaFuente, E.J.; Liu, W.C.V. J. Toxicol. Environ. Health 1985, 16, pp 449-460. 18. Groves, J.T.; McClusky, G.A.; White, R.E.; Coon, M.J. Biochem. Biophys. Res. Commun. 1978, 81, pp 154-160. 19. White, R.E.; Groves, J.T.; McClusky, G.A. Acta Biol. Med. Ger. 1979, 38, pp 475-182. 20. Guengerich, P.F.; MacDonald, T.L. Acc. Chem. Res. 1984, 17, pp 9-16. 21. DeVito, S.C.; Pearlman, R.S. Med. Chem. Res. 1992, 1, pp 461-465. 22. Grogan, J.; DeVito, S.C.; Pearlman, R.S.; Korzekwa, K.R. Chem. Res. Toxicol. 1992, 5, pp 548-552. 23. Estabrook, R.W.; Werringloer, J. In ACS Symposium Series, 44th Ed., D.M. Jerina, Ed. American Chemical Society: Washington DC, 1977, pp 1. 24. Korshunov, Y . N . Gig Primen., Toksikol. Pestits. Klin. Otravlenii, 1970, No. 8, pp 398-403 (see Chemical Abstracts 77:122773). 25. Parkinson, A . In Cassarett and Doull's Toxicology, the Basic Science of Poisons, Fifth Edition. Klaassen, C.D.; McGraw-Hill: New York, 1996, pp 113-186. 26. Mumtaz, M . M . ; Farooqui, M.Y.H.; Ghanayem, B.I.; Ahmed, A . E . Toxicol. Appl. Pharmacol. 1991, 110, pp 61-69. 27. Keiser, H.R.; Harris, E.D.; Sjoerdsma, A . Clin. Pharm. Ther. 1967, 8, pp 587-592. 28. Registry of Toxic Effects of Chemical Substances (on-line version).



10.

DEVTTO


221

29. EPA document 8(e) HQ-0988-0754. 30. Dangerous Properties of Industrial Materials , 7th Edition, Vol. II, Sax, N.I.; Lewis, R.J., Eds., Van Nostrand Reinhold: New York, N.Y., 1989. 31. Smyth, H.F.; Carpenter,C.P.;Weil, C.S.; Urbano, P.C.; Striegel, J.A. Am. Ind. Hyg. Assoc. J. 1962, 23, pp 95-107. 32. EPA document OTS-50-8001950 (Microfiche No. OTS 205790; P-80-311) 33. EPA document 8(e)HQ-0988-0754 34. Registry of Toxic Effects of Chemical Substances, 1985-86, Sweet, D.V., Ed.; Government Printing Office, Washington, D.C., 1988. 35. Stamler, F.W. Exp. Biol. and Med. 1955, 90, pp 294-298. 36. Kalliomaki, L.; Yli-Pohja, M . ; Kulonen, E. Experientia 1957, 13, p 495. 37. Dasler, W. J. Nutr. 1954, 53, pp 105-113. 38. Selye, H . Revue Canadienne De Biologie 1957, 16, pp 1-82. 39. Barrows, M.V.; Simpson, C.F.; Miller, E.J. Quart. Rev. Biol. 1974, 49, pp 101-128. 40. Roy, D.N. Nutr. Abstr. Rev. (Ser. A). 1981, 51, pp 691-706. 41. Schilling, E.D.; Strong, F.M. J. Am. Chem. Soc. 1954, 76, p 2848. 42. Dasler, W. Science 1954, 120, pp 307-308. 43. Schilling, E.D.; Strong, F.M. J. Am. Chem. Soc. 1955, 77, pp 2843-2845. 44. Geiger, B.J.; Steenbock, H.; Parsons, H.T. J. Nutr. 1933, 6, pp 427-442. 45. Ponseti, I.V.; Baird, W.A. Amer.J.Pathol. 1952, 28, pp 1059-1078. 46. Ponseti, I.V.; Shepard, R.S. J. Bone and Joint Surg. 1954, 36a, pp 1031-1058. 47. Bachhuber, T.E.; Lalich, J.J.; Angevine, D.M.; Schilling, E.D.; Strong, , F.M. Proc. Soc. Exp. Biol. Med. 1955, 89, pp 294-297. 48. Lalich, J.J. Arch. Path. 1956, 61, pp 520-524. 49. Clemmons, J.J.; Angevine, D.M. Am. J. Path. 1957, 33, pp 175-187. 50. Julian, M . ; Pieraggi, M . TH.; Bouissou, H. Pharmacol. Res. Commun. 1979, 11, pp 501-508. 51. Simpson, C.F. Exp. Molec. Pathol. 1982, 37, pp 382-392. 52. Simpson, C.F.; Cardeilhac, P.T. Proc. Soc. Exp. Biol. Med. .1983, 172, pp 168-172. 53. Wawzonek, S.; Ponseti, I.V.; Shepard, R.S.; Wiedenmann, L . G . Science. 1955, 121, pp 63-65. 54. Levene, C.I. J. Exp. Med. 1961, 114, pp 295-310. 55. Roy, D.N.; Lipton, S.H.; Bird, H.R.; Strong, F.M. Proc. Soc. Exp. Biol. Med. 1960, 103, pp 286-289. 56. Levene, C.I.; Carrington, M.J. Biochem. J. 1985, 232, pp 293-296. 57. Ronchetti, I.P.; Contri, M.B.; Fornieri, C.; Quaglino Jr., D.; Mori, G. Connect. Tissue Res. 1985, 14, pp 159-167. 58. Wilmarth, K.R.; Froines, J.R. J. Toxicol. Environ. Health 1992, 37, pp 411-423. 59. Levene, C.I. Fed. Proc. 1958, 17, p 445.


222 60. 61. 62. 63. 64. 65. Downloaded by UNIV OF CALIFORNIA SAN DIEGO on August 27, 2015 | http://pubs.acs.org Publication Date: May 5, 1996 | doi: 10.1021/bk-1996-0640.ch010

66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90.


Levene, C.I.; Gross, J. Fed. Proc. 1959, 18, p 90. Gross, J.; Levene, C.I. Am. J. Path. 1959, 35, pp 687-688. Levene, C.I.; Gross, J. J. Exp. Med. 1959, 110, pp 771-790. Martin, G.R.; Piez, K.A.; Lewis, M.S. Biochim. Biophys. Acta 1963, 69, pp 472-479. Davies, I.; Schofield, D. Exp. Geront. 1980, 15, pp 487-494. O'Dell, B.L.; Elsden, D.F.; Thomas, J.; Partridge, S.M. Nature. 1966, 209, pp 401-402. Page, R.C.; Benditt, E.P. Proc. Soc. Exp. Biol. Med. 1967, 124, pp 454-459. Lalich, J.J. Science. 1958, 128, pp 206-207. Page, R.C.; Benditt, E.P. Proc. Soc. Exp. Biol. Med. 1967, 124, pp 459-165. Narayanan, A.S.; Siegel, R.C.; Martin, G.R. Biochim. Biophys. Res. Commun. 1972, 46, pp 745-751. Page, R.C.; Benditt, E.P. Lab. Inv. 1972, 26, pp 22-26. Chvapil, M . ; Misiorowski, R.; Eskelson, C. J. Surg. Res. 1981, 31, pp 151-155. Fleisher, J.H.; Speer, D.; Brendel, K.; Chvapil, M.Toxicol. Appl. Pharmacol. 1979, 47, pp 61-69. Rando, R.R. Acc. Chem. Res. 1975, 8, pp 281-288. Clemmons, J.J. Fed. Proc. 1958, 17, p 432. Schmidt, W.; Wendler, D. Folia Morphol 1982, 30, pp 146-148. Maycock, A.L.; Suva, R.H.; Abeles, R.H. J. Am. Chem. Soc. 1975, 97, pp 5613-5614. Pinnell, S.R.; Martin, G.R. Proc. Nat. Acad. Sci. U.S., 1968, 61, pp 708-716. Piez, K.A. Ann. Rev. Biochem. 1968, 37, pp 547-570. Gallop, P.M.; Blumenfield, O.O.; Seifter, S. Ann. Rev. Biochem. 1972, 41, pp 617-672. Bailey, A.J.; Robins, S.P. Sci. Prog. 1976, 63, pp 419-444. Bailey, A.J.; Robins, S.P.; Balian, G. Nature. 1974, 252, pp 105-110. Sandberg, L.B. Int. Rev. Connect. Tissue Res. 1976, 7, pp 159-210. Bailey, A.J. J. Clin. Path. (Suppl.) 1978, 12, pp 49-58. Bornstein, P.; Piez, K.A. Biochemistry 1966, 5, pp 3460-3473. Bornstein, P.; Kang, A.H.; Piez, K.A. Biochemistry. 1966, 5, pp 3803-3812. Bornstein, P.; Kang, A.H.; Piez, K.A. Proc. Nat. Acad. Sci. U.S. 1966, 55, pp 417-424. Piez, K.A.; Martin, G.R.; Kang, A.H.; Bornstein, P.Biochemistry. 1966, 5, pp 3813-3820. Kang, A.H.; Bornstein, P.; Piez, K.A.Biochemistry. 1967, 6, pp 788-795. Lent, R.; Franzblau, C.Biochem. Biophys. Res. Commun. 1967, 26, pp 43-50. Miller, E.J.; Martin, G.R.; Piez, K . A . Biochem. Biophys. Res. Commun. 1964, 17, pp 248-253.



10.

DEVITO


223

91. Anwar, R.A.; Oda, G. Biochim. Biophys. Acta. 1967, 133, pp 151-156. 92. Pinnell, S.R.; Martin, G.R. Proc. Nat. Acad. Sci. U.S. 1968, 61, pp 708-716. 93. Tang, S-S.; Trackman, P.C.; Kagan, H.M. J. Biol. Chem. 1983, 258, pp 4331-4338. 94. Tanii, H.; Hayashi, M.; Hashimoto, K. Neurotoxicology 1989, 10, pp 157-166. 95. ibid. Arch. Toxicol. 1990, 64, pp 231-236. 96. Pestronk, A.; Keogh, J.P.; Griffin, J.W.Neurology 1979, 29, p 540. 97. Kreiss, K.; Wegman, D.H.; Niles, C.A.; Siroky, M.B.; Krane, R.J.; Feldman, R.G. JAMA 1980, 243, pp 741-745. 98. Keogh, J.P.; Pestronk, A.; Wertheimer, D.; Moreland, R. JAMA 1980, 243, pp 746-749. 99. Baker, E.L.; Christiani, D.C.; Wegman, D.H.; Siroky, M.; Niles, C.A.; Feldman, R.G. Scand. J. Work Environ. Health. 1981, 70(Suppl. 4), pp 54-59. 100. Gad, S.C.; McKelvey, J.A.; Turney, R.A. Drug and Chem. Toxicol. 1979, 2, pp 223-236. 101. Jaeger, R.J.; Plugge, H. J. Environ. Pathol. Toxicol. 1980, 4, pp 555-562. 102. Froines, J.R. Biochem. Pharmacol. 1985, 34, pp 3185-3187. 103. Froines, J.R.; Watson, A.J.Toxicol. Environ. Health 1985, 16, pp 469-179. 104. Wilmarth, K.R.; Froines, J.R. J. Toxicol. Environ. Health 1991, 32, pp 415-427. 105. Ahmed, A.E.; Farooqui, M.Y.H. Pharmacologist 1984, 26, p 199. 106. Mumtaz, M.M.; Farooqui, M.Y.H.; Ghanayem, B.I.; Ahmed, A.E. Toxicol. Appl. Pharmacol. 1991, 110, pp 61-69. 107. Hartmann, H.A.; Stich, H.F. Fed. Proc. 1957, 16, pp 358-359. 108. Peele, D.B.; Allison, S.D.; Crofton, K.M. Toxicol. Appl. Pharmacol. 1990,105, pp 321-332. 109. Schulze, G.E.; Boysen, B.G. Fundam. Appl. Toxicol. 1991, 16, pp 602-615. 110. Crofton,K.M.;Knight, T. Neurotoxicol. and Teratology 1991, 13, pp 575-581. 111. Genter, M.B.; Llorens, J.; O'Callaghan, J.P.; Peele, D.B.; Morgan, K.T.; Crofton, K M . J. Pharmacol, and Experim. Therap. 1992, 263, pp 1432-1439. 112. Goldey, E.S.; Kehn, L.S.; Crofton, K M . Hearing Research 1993, 69, pp 221-228. 113. Llorens, J.; Dememes, D.; Sans, A. Toxicol. Appl. Pharmacol. 1993, 123, pp 199-210. 114. Llorens, J.; Crofton, K M . ; O'Callaghan, J.P. J. Pharmacol. and Experim. Therap. 1993, 265, pp 1492-1498. 115. Genter, M.B.; Deamer, N.J.; Cao, Y.; Levi, P.E. J. Biochem. Toxicol 1994, 9, pp 31-39. 116. Morandi, A.; Gambetti, P.; Arora, P.K.; Sayre, L.M.; Brain Research 1987, 437, pp 69-76. 117. Jacobson, A.R.; Coffin, S.H.; Shearson, C.M.; Sayre, L . M . Molecular Toxicol. 1987, 1, pp 17-34.


Designing Safer Nitriles - ACS Symposium Series (ACS Publications)

Recommend Documents